Hello everyone,
Can you help me with some explanation regarding FMOLS on panel in EViews?
I am not very sure what to choose in the case of trend specification: None, Constant (at level), Linear trend….and also for Additional trends: None, Linear trend or Quadratic. Also for the panel method: pooled, pooled (weighted), or grouped?
For my panel, I have 11 countries with 24 years. Thank you!
The following link might be helpful to you:
http://www.eviews.com/help/helpintro.html#page/content/pancoint-Estimating_Panel_Cointegration_in_EViews.html
I suggest and hope the readings on panel FMOLS can also help you:
Phillips & Hansen (1990), Phillips & Moon (1999), Pedroni (2000), and Kao & Chiang (2000).
Dear Mrs. Eugenia Ramona Mara
I hope by reading the references from Mr. Elvis Munyaradzi and Mr. Syed Ali Raza have helped you. But here I also add some references to help you :
Baltagi, Badi and Chihwa Kao (2000). “Nonstationary Panels, Cointegration in Panels and Dynamic Panels: A Survey,” in Baltagi, B. H. ed., Nonstationary Panels, Panel Cointegration and Dynamic Panels, 15, Amsterdam: Elsevier, 7–51.
Breitung, Jörg and M. Hashem Pesaran (2008). “Unit Roots and Cointegration in Panels,” in Mátyás, László and Patrick Sevestre, eds. The Econometrics of Panel Data, Berlin: Springer-Verlag Berlin Heidelberg.
Dear Eugenia.
I recommend you read through Phillips and Hansen (1990) and Pedroni (2000) for more information.
Regards
Dear colleague,
Very reasonable question.
Concerning the choice of trend this is helped graphically.
As you mentioned for FMOLS (and DOLS) method(s) there are three different estimators: pooled, pooled weighted and grouped mean.
The first extens Phillips and Hansen (1990) FMOLS estimator from Phillips and Moon (1999) after correcting deterministic components in regressand and regressors.
If you want to allow different LR variances across cross-sections and for heterogeneous panels Pedroni (2000) and Kao and Chiang (2000) suggest pooled weighted FMOLS.
While grouped mean FMOLS (Pedroni 2000, 2001) relies on averaging individual cross-section FMOLS estimates.
I hope it helps. All proposed references from the answers so far are good.
Dear Eugenia
Apart from the main question and before all another important issue before is the cross sectional dependence
Dear George,
yes indeed, I already tested the Cross sectional dependence, and because was confirmed I applied PCSE. And in the next stage, I used FMOLS.
Dear George and Eugenia,
What is the role of presence/absence of cross-sectional dependence on deciding to choose FMOLS ? choice of FMOLS as a model depends on state of cross-sectional dependence?
Kind Regards
Vusal
Dear Vusal,
I used FMOLS because I also have cointegration, I think there is no relation between cross-section dependence and FMOLS, but I suppose if you detect cross-section dependence is necessary to verify also cointegration. Maybe other researchers specialized in econometrics can come up with more information :)
Kind regards,
Ramona
- Badges
- Science topic